CN106203345B - An image control device with identity verification function - Google Patents

Abstract

The present invention provides an image control device with identity authentication function, comprising an image control device and an iris recognizer connected to the image control device by electrical signals, wherein the iris recognizer comprises: (1) a sampling module; (2) a preprocessing module; (3) a feature encoding module for extracting and encoding features of an iris image, comprising a first LBP operator processing submodule, a second LBP operator processing submodule, a third LBP operator processing submodule and a fourth LBP operator processing submodule; (4) an encoding matching module. The present invention increases the correlation between a center point and other surrounding neighborhoods, and can satisfy image textures of different scales and frequencies. After being processed by multiple LBP operator processing submodules, the encoding length is continuously reduced without affecting the correlation between the center point and the surrounding neighborhoods, thereby saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining a higher robustness.

Description

An image control device with identity verification function

technical field

The invention relates to the field of image control device system design, in particular to an image control device with an identity verification function.

Background technique

In the related art, an image control device with an identity verification function usually uses a basic LBP (local binary pattern) operator to extract and encode the iris image features. The LBP operator is a method for describing the texture features in the grayscale range of the image. It has strong robustness to illumination changes, so it is widely used in image texture feature extraction.

The basic LBP operator is generally defined as: in a 3×3 window, it consists of the center point n _c and its surrounding 8 neighborhoods n ₀ ,...n ₇ , where the texture T is defined as: T=(n _0- n _c ,n _1- n _c ,...,n _7- n _c ), binarize it, take n _c as the threshold, compare the 8 points in the neighborhood with n _c , if it is greater than the value of the center point, mark is 1, otherwise it is marked as 0. The binarized texture T is: T=(sgn(n ₀ -n _c ),sgn(n ₁ -n _c ),...,sgn(n ₇ -n _c )), where After calculation, 8 binary numbers with n _c as the center will be obtained, and then the weighted summation of different pixel positions will obtain the LBP value of the center point, where the calculation formula of the LBP value is: LBP operation is performed on each pixel in the image to obtain the LBP texture description of the image.

However, since the basic LBP operator only covers 8 neighborhood pixels of the center point, its correlation with other surrounding neighborhoods is not comprehensive enough to satisfy image textures of different scales and frequencies.

SUMMARY OF THE INVENTION

In view of the above problems, the present invention provides an image control device with an identity verification function with fast recognition speed and wide recognition range, and solves the problem of the image control device system in the related art that uses the basic LBP operator to extract and encode the features of the iris image. It cannot handle the problem of image textures of different scales and frequencies.

The object of the present invention adopts the following technical solutions to realize:

An image control device with an identity verification function, comprising an image control device and an iris identifier connected to the image control device with an electrical signal, the image control device comprising:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to acquire and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The image is subjected to plane correction, and an image correction sub-module is set, and the correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

(2) a preprocessing module for positioning and normalizing the acquired iris image, including a light spot filling sub-module, which is used for each light spot detected in the iris image For filling, the gray value of the four envelope points in the non-spot area adjacent to the light spot is used to calculate the gray value of the light spot, and a light spot in the iris image is defined as P ₀ (x ₀ , y ₀ ), the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ (x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ ) in sequence , y ₄ ), the calculation formula of the gray value of the defined light spot is:

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with K pixels in the 5×5 window to calculate the LBP value, and the K pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the K pixels are marked as n ₀ to n _K , the value range of K is [20, 24], and the value range of 1st-LBP(x _c , y _c ) is [0, K];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the feature encoding length of the rectangular image processed by the second LBP operator processing sub-module, which is centered on point n _c in a 3×3 window According to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i=0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: it is used to continue to reduce the coding length on the basis of the third LBP operator processing sub-module processing. The calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module further includes:

(1) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. During cutting, the pupil position is taken as the center and a radius of 5 times is used to fill the iris after the light spot. image to be cut;

(2) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(3) The normalization submodule is used to expand the iris region after positioning into a fixed resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

The beneficial effects of the present invention are:

1. Set the image correction sub-module, and define the correction formula, which improves the accuracy of image processing;

2. Set the light spot filling sub-module, and define the gray value calculation formula of the light spot, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned;

3. The set initial positioning unit, which locates the inner and outer circles of the iris through the improved Canny edge detection operator and Hough circle detection, which facilitates the positioning of the iris and improves the speed of the iris;

4. The first LBP operator processing sub-module is set, which increases the correlation between the center point and other surrounding neighborhoods, and can satisfy image textures of different scales and frequencies;

5. The set second LBP operator processing sub-module, third LBP operator processing sub-module and fourth LBP operator processing sub-module are continuously reduced without affecting the correlation between the center point and the surrounding neighborhood. The coding length saves storage space, reduces the amount of calculation, improves the recognition speed, enhances the recognition accuracy, and obtains higher robustness.

Description of drawings

The present invention will be further described by using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, under the premise of no creative work, other Attached.

FIG. 1 is a schematic diagram of the connection of the iris recognizer of the present invention.

FIG. 2 is a schematic diagram of an image control device of the present invention.

Detailed ways

The present invention will be further described with reference to the following examples.

Example 1

Referring to FIG. 1 and FIG. 2 , an image control device with an identity verification function in this embodiment includes an image control device and an iris recognizer electrically connected to the image control device. The image control device includes:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to obtain the iris image and collect the information of the iris image;

(2) The preprocessing module is used to obtain and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The iris image is plane corrected, and an image correction sub-module is set. The correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with 20 pixels in a 5×5 window to calculate the LBP value, and the 20 pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the 20 pixel points are marked as n ₀ -n ₂₀ , and the value range of 1st-LBP(x _c , y _c ) is [0, 20];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the processing time of the second LBP operator processing sub-module

rectified rectangle

The feature encoding length of the image, centered on point n _c , in a 3×3 window according to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i= 0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: used after the third LBP operator processing sub-module processing

on the basis of

Continue to reduce the encoding length, the calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module includes:

(1) Light spot filling sub-module: It is used to fill each light spot detected in the iris image, and the gray value of the four envelope points in the non-spot area adjacent to the light spot is used for filling. To calculate the gray value of the light spot, define a light spot in the iris image as P ₀ (x ₀ , y ₀ ), and the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ ( x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ , y ₄ ), the calculation formula of the gray value of the defined light spot is:

(2) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. When cutting, take the pupil position as the center and a radius of 5 times to fill the iris after the light spot. image to be cut;

(3) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(4) The normalization sub-module is used to expand the positioned iris region into a fixed-resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

In this embodiment, the light spot filling sub-module is set, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned; the initial positioning unit is set, which uses the improved Canny edge detection operator and Hough circle. The detection locates the inner and outer circles of the iris, which facilitates the positioning of the iris and improves the speed of the iris; the first LBP operator processing sub-module is set to increase the correlation between the center point and other surrounding neighborhoods, which can meet the needs of different scales and frequencies. image texture; set the second LBP operator processing sub-module, the third LBP operator processing sub-module and the fourth LBP operator processing sub-module, without affecting the correlation between the center point and the surrounding neighborhood, Continuously reducing the code length, saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining high robustness, when using the CASIAV1.0 iris library for testing, the results are as follows:

Example 2

Referring to FIG. 1 and FIG. 2 , an image control device with an identity verification function in this embodiment includes an image control device and an iris recognizer electrically connected to the image control device. The image control device includes:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to obtain the iris image and collect the information of the iris image;

(2) The preprocessing module is used to obtain and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The iris image is plane corrected, and an image correction sub-module is set. The correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with 21 pixels in the 5×5 window to calculate the LBP value, and the 21 pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the 21 pixel points are marked as n ₀ to n ₂₁ , and the value range of 1st-LBP(x _c , y _c ) is [0, 21];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the feature encoding length of the rectangular image processed by the second LBP operator processing sub-module, which is centered on point n _c in a 3×3 window According to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i=0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: it is used to continue to reduce the coding length on the basis of the third LBP operator processing sub-module processing. The calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module includes:

(1) Light spot filling sub-module: It is used to fill each light spot detected in the iris image, and the gray value of the four envelope points in the non-spot area adjacent to the light spot is used for filling. To calculate the gray value of the light spot, define a light spot in the iris image as P ₀ (x ₀ , y ₀ ), and the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ ( x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ , y ₄ ), the calculation formula of the gray value of the defined light spot is:

(2) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. When cutting, take the pupil position as the center and a radius of 5 times to fill the iris after the light spot. image to be cut;

(3) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(4) The normalization sub-module is used to expand the positioned iris region into a fixed-resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

In this embodiment, the light spot filling sub-module is set, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned; the initial positioning unit is set, which uses the improved Canny edge detection operator and Hough circle. The detection locates the inner and outer circles of the iris, which facilitates the positioning of the iris and improves the speed of the iris; the first LBP operator processing sub-module is set to increase the correlation between the center point and other surrounding neighborhoods, which can meet the needs of different scales and frequencies. image texture; set the second LBP operator processing sub-module, the third LBP operator processing sub-module and the fourth LBP operator processing sub-module, without affecting the correlation between the center point and the surrounding neighborhood, Continuously reducing the code length, saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining high robustness, when using the CASIAV1.0 iris library for testing, the results are as follows:

Example 3

Referring to FIG. 1 and FIG. 2 , an image control device with an identity verification function in this embodiment includes an image control device and an iris recognizer electrically connected to the image control device. The image control device includes:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to acquire and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The image is subjected to plane correction, and an image correction sub-module is set, and the correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

(2) The preprocessing module is used to locate and normalize the acquired iris image;

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with 22 pixels in the 5×5 window to calculate the LBP value, and the 22 pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the 22 pixels are marked as n ₀ to n ₂₁ , and the value range of 1st-LBP(x _c , y _c ) is [0, 22];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the feature encoding length of the rectangular image processed by the second LBP operator processing sub-module, which is centered on point n _c in a 3×3 window According to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i=0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: it is used to continue to reduce the coding length on the basis of the third LBP operator processing sub-module processing. The calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module includes:

(1) Light spot filling sub-module: It is used to fill each light spot detected in the iris image, and the gray value of the four envelope points in the non-spot area adjacent to the light spot is used for filling. To calculate the gray value of the light spot, define a light spot in the iris image as P ₀ (x ₀ , y ₀ ), and the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ ( x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ , y ₄ ), the calculation formula of the gray value of the defined light spot is:

(2) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. When cutting, take the pupil position as the center and a radius of 5 times to fill the iris after the light spot. image to be cut;

(3) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(4) The normalization sub-module is used to expand the positioned iris region into a fixed-resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

In this embodiment, the light spot filling sub-module is set, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned; the initial positioning unit is set, which uses the improved Canny edge detection operator and Hough circle. The detection locates the inner and outer circles of the iris, which facilitates the positioning of the iris and improves the speed of the iris; the first LBP operator processing sub-module is set to increase the correlation between the center point and other surrounding neighborhoods, which can meet different scales and frequencies. image texture; set the second LBP operator processing sub-module, the third LBP operator processing sub-module and the fourth LBP operator processing sub-module, without affecting the correlation between the center point and the surrounding neighborhood, Continuously reducing the coding length, saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining high robustness, when using the CASIAV1.0 iris library for testing, the results are as follows:

Example 4

Referring to FIG. 1 and FIG. 2 , an image control device with an identity verification function in this embodiment includes an image control device and an iris recognizer electrically connected to the image control device. The image control device includes:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to acquire and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The image is subjected to plane correction, and an image correction sub-module is set, and the correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

(2) The preprocessing module is used to locate and normalize the acquired iris image;

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with 23 pixels in the 5×5 window to calculate the LBP value, and the 23 pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the 23 pixel points are marked as n ₀ to n ₂₁ , and the value range of 1st-LBP(x _c , y _c ) is [0, 23];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the feature encoding length of the rectangular image processed by the second LBP operator processing sub-module, which is centered on point n _c in a 3×3 window According to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i=0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: it is used to continue to reduce the coding length on the basis of the third LBP operator processing sub-module processing. The calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module includes:

(1) Light spot filling sub-module: It is used to fill each light spot detected in the iris image, and the gray value of the four envelope points in the non-spot area adjacent to the light spot is used for filling. To calculate the gray value of the light spot, define a light spot in the iris image as P ₀ (x ₀ , y ₀ ), and the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ ( x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ , y ₄ ), the calculation formula of the gray value of the defined light spot is:

(2) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. When cutting, take the pupil position as the center and a radius of 5 times to fill the iris after the light spot. image to be cut;

(3) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(4) The normalization sub-module is used to expand the positioned iris region into a fixed-resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

In this embodiment, the light spot filling sub-module is set, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned; the initial positioning unit is set, which uses the improved Canny edge detection operator and Hough circle. The detection locates the inner and outer circles of the iris, which facilitates the positioning of the iris and improves the speed of the iris; the first LBP operator processing sub-module is set to increase the correlation between the center point and other surrounding neighborhoods, which can meet the needs of different scales and frequencies. image texture; set the second LBP operator processing sub-module, the third LBP operator processing sub-module and the fourth LBP operator processing sub-module, without affecting the correlation between the center point and the surrounding neighborhood, Continuously reducing the code length, saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining high robustness, when using the CASIAV1.0 iris library for testing, the results are as follows:

Example 5

Referring to FIG. 1 and FIG. 2 , an image control device with an identity verification function in this embodiment includes an image control device and an iris recognizer electrically connected to the image control device. The image control device includes:

a first decoder for decoding the input video data stream;

an encoder for encoding the output of the first decoder at a plurality of converted bit rates;

an output unit for outputting the image data encoded by the encoder at the first bit rate for recording;

a second decoder for decoding the same image data encoded by the encoder at the second bit rate;

a control unit for determining the second bit rate based on the first bit rate.

Preferably, it is characterized in that the sum of the first bit rate and the second bit rate is less than the input bit rate.

Preferably, the image control device further comprises:

an encryption unit, configured to encrypt the input image data stream;

The storage unit is used to store the encrypted image data.

Preferably, it is characterized in that, described iris recognizer comprises:

(1) Sampling module, used to acquire and correct the iris image and collect the information of the iris image. Since the actual obtained iris image and the standard collected iris image will slightly deviate on the same plane, it is necessary to analyze the actual obtained iris image. The image is subjected to plane correction, and an image correction sub-module is set, and the correction formula adopted by the image correction sub-module is:

Among them, I(x,y) _A represents the actual obtained iris image, I(x,y) _B represents the standard collected iris image, the standard value between the actual obtained iris image and the standard collected iris image pixel value Difference;

(2) The preprocessing module is used to locate and normalize the acquired iris image;

Preferably, it is characterized in that the iris identifier further comprises:

(3) The feature encoding module is used to extract and encode the features of the iris image, including:

a. The first LBP operator processing sub-module: used to compare any point n _c in the iris image with 24 pixels in the 5×5 window to calculate the LBP value, and the 24 pixels are represented by point n _c is the center distributed on the periphery of point n _c , and the coordinates of n _c are set as (x _c , y _c ), the calculation formula of the LBP value is:

Wherein, the 24 pixel points are marked as n ₀ to n ₂₁ , and the value range of 1st-LBP(x _c , y _c ) is [0, 24];

b. The second LBP operator processing sub-module is used to strengthen the correlation between the point n _c and the surrounding neighborhood on the premise of ensuring the coding length, and it takes the 8 neighborhood pixels of the point n _c as the sub-center Points, denoted as n _vc0 ,n _vc1 ,...,n _vc7 , use a 3×3 window, and use the mean value of all pixels in the window Instead of the value of the sub-center point, use the LBP operator to calculate the center point n _c . The calculation formula is:

c. The third LBP operator processing sub-module is used to shorten the feature encoding length of the rectangular image processed by the second LBP operator processing sub-module, which is centered on point n _c in a 3×3 window According to the custom function {n _vcj ,|n _vcj -n _c |=rank ₄ (|n _vci -n _c |,i=0,1,...,7),j=0,1,2,3} Select 4 sub-center points for calculation, the calculation formula is:

Among them, rank ₄ (|n _vci -n _c |, i=0,1,...,7) indicates that the values of 7 |n _vci -n _c | are arranged from small to large and the first 4 numbers are taken, n _vcj represents the selected 4 sub-center points;

d. The fourth LBP operator processing sub-module: it is used to continue to reduce the coding length on the basis of the third LBP operator processing sub-module processing. The calculation formula is:

After the calculation is completed, the code representing the iris image feature is output;

(4) The code matching module is used to receive the code representing the iris image feature and compare it with the feature code in the database to complete the identification of the identity.

Wherein, the preprocessing module includes:

(1) Light spot filling sub-module: It is used to fill each light spot detected in the iris image, and the gray value of the four envelope points in the non-spot area adjacent to the light spot is used for filling. To calculate the gray value of the light spot, define a light spot in the iris image as P ₀ (x ₀ , y ₀ ), and the four envelope points are P ₁ (x ₁ , y ₁ ), P ₂ ( x ₂ , y ₂ ), P ₃ (x ₃ , y ₃ ), P ₄ (x ₄ , y ₄ ), the calculation formula of the gray value of the defined light spot is:

(2) Coarse positioning sub-module: connected with the light spot filling sub-module, used to cut the iris image and preliminarily locate the pupil position. When cutting, take the pupil position as the center and a radius of 5 times to fill the iris after the light spot. image to be cut;

(3) Fine positioning sub-module: connected with the coarse positioning sub-module for precise positioning of the iris area;

(4) The normalization sub-module is used to expand the positioned iris region into a fixed-resolution iris image.

Wherein, the precise positioning sub-module includes a downsampling unit, a primary positioning unit and a repositioning unit connected in sequence, the downsampling unit is used for downsampling the cut iris image, and the primary positioning unit is used for improving The Canny edge detection operator and the Hough circle detection are used to locate the inner and outer circles of the iris.

The improved Canny edge detection operator is a Canny edge detection operator that only suppresses non-maximum values in the vertical direction.

The improved Canny edge detection operator is a Canny edge detection operator that only uses a high threshold for strong edge detection.

In this embodiment, the light spot filling sub-module is set, which well preserves the structural information of the iris image, and the filled iris image can be effectively positioned; the initial positioning unit is set, which uses the improved Canny edge detection operator and Hough circle. The detection locates the inner and outer circles of the iris, which facilitates the positioning of the iris and improves the speed of the iris; the first LBP operator processing sub-module is set to increase the correlation between the center point and other surrounding neighborhoods, which can meet the needs of different scales and frequencies. image texture; set the second LBP operator processing sub-module, the third LBP operator processing sub-module and the fourth LBP operator processing sub-module, without affecting the correlation between the center point and the surrounding neighborhood, Continuously reducing the code length, saving storage space, reducing the amount of calculation, improving the recognition speed, enhancing the recognition accuracy, and obtaining high robustness, when using the CASIAV1.0 iris library for testing, the results are as follows:

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, not to limit the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that , the technical solutions of the present invention may be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims (5)

1. a kind of image control device with identification verification function, including image control device and with image control device telecommunications The iris recognition device of number connection, the image control device includes:

First decoder, for being decoded to input image data flow；

Encoder, for being encoded with the bit rate after multiple conversions to the output of first decoder；

Output unit, for exporting the encoder with the image data of the first bit rate coding, to be recorded；

Second decoder, for being decoded to the encoder with the same image data of the second bit rate coding；

Control unit, for determining second bit rate based on first bit rate；

The summation of first bit rate and second bit rate is less than the input bit rate；

The image control device further include:

Encryption unit, for being encrypted to the input image data flow；

Storage unit, for storing encrypted image data；

The iris recognition device includes:

(1) sampling module, for obtaining, correcting iris image and acquire the information of iris image, due to the iris actually obtained Can slightly have deviation between image and the iris image of standard acquisition in the same plane, need to the iris image actually obtained Plane correction is carried out, image rectification submodule, the updating formula that described image correction module uses are set are as follows:

Wherein, I (x, y)_AIndicate the iris image actually obtained, I (x, y)_BIndicate the iris image of standard acquisition, it is practical to obtain Iris image and standard acquisition iris image each pixel point value between standard deviation；

(2) preprocessing module, for carrying out positioning and normalized to the iris image of acquisition comprising hot spot point filling Module, hot spot point filling submodule is for being filled each hot spot point detected in iris image, benefit when filling The gray value of hot spot point is calculated with the gray value of four envelope points up and down in the non-spot area adjacent with hot spot point, Defining a hot spot point in iris image is P₀(x₀,y₀), four envelopes point is followed successively by P₁(x₁,y₁)、P₂(x₂,y₂)、P₃ (x₃,y₃)、P₄(x₄,y₄), define the gray value calculation formula of hot spot point are as follows:

；

The iris recognition device further include:

(3) feature coding module is extracted and is encoded for the feature to iris image, comprising:

A, first time LBP operator handles submodule: for any point n in iris image_cWith K pixel in 5 × 5 windows Point is compared to calculate LBP value, and the K pixel is with point n_cCentered on be distributed in point n_cPeriphery, if n_cCoordinate be (x_c, y_c), the calculation formula of LBP value are as follows:

Wherein, the K pixel is labeled as n₀~n_K, the value range of K is [20,24], 1st-LBP (x_c,y_c) value model It encloses for [0, K]；

B, second of LBP operator handles submodule, for reinforcing the point n under the premise of guaranteeing code length_cWith surrounding neighbors Relevance, with point n_c8 neighborhood territory pixel points as sub-center point, be denoted as n_vc0,n_vc1,...,n_vc7, using 3 × 3 windows, With the mean value of entire pixels in windowInstead of the value of sub-center point, LBP operator is reused to central point n_cIt carries out It calculates, calculation formula are as follows:

C, third time LBP operator handles submodule, for shortening through second LBP operator processing submodule treated iris figure The feature coding length of picture, with point n_cCentered on, according to custom function { n in 3 × 3 window_vcj,|n_vcj-n_c|= rank₄(n_vci-n_c|, i=0,1 ..., 7), j=0,1,2,3 } 4 sub-center points of selection are calculated, calculation formula are as follows:

Wherein, rank₄(|n_vci-n_c|, i=0,1 ..., 7) indicate to 7 | n_vci-n_c| value carry out from small to large arrange after take Preceding 4 numbers, n_vcjIndicate the 4 sub-center points chosen；

D, the 4th LBP operator handles submodule: for continuing on the basis of treated in third time LBP operator processing submodule Reduce code length, calculation formula are as follows:

Output indicates the coding of iris image feature after having been calculated；

(4) codes match module, for receiving the coding for indicating iris image feature and by itself and the feature in database Coding is compared, and completes the identification to identity.

2. a kind of image control device with identification verification function according to claim 1, characterized in that the pre- place Manage module further include:

(1) coarse positioning submodule: connecting with hot spot point filling submodule, for cut simultaneously Primary Location pupil to iris image Hole site, when cutting centered on the pupil position, 5 times of radius cuts the iris image after filling hot spot；

(2) it fine positioning submodule: is connect with coarse positioning submodule, for being accurately positioned iris region；

(3) submodule is normalized, for the iris region after positioning to be launched into the iris image of fixed resolution.

3. a kind of image control device with identification verification function according to claim 2, characterized in that the essence is fixed Bit submodule includes sequentially connected downsampling unit, first positioning unit and repositions unit, and the downsampling unit is used In carrying out down-sampling to the iris image after cutting, the first positioning unit is used to calculate by improved Canny edge detection Son and Hough loop truss position iris inside and outside circle, described to reposition what unit was used to position with first positioning unit Parameter is accurately positioned on iris image.

4. a kind of image control device with identification verification function according to claim 3, characterized in that the improvement Canny edge detection operator be only vertical direction is carried out non-maximum inhibition a Canny edge detection operator.

5. a kind of image control device with identification verification function according to claim 4, characterized in that the improvement Canny edge detection operator be only with high threshold carry out strong edge detection Canny edge detection operator.